zoot

Slashdot: From the proverbial horse's mouth here.

I can't immediately imagine how that would work. Maybe a brute-force way could be to interpolate a smooth path and just let the chips fall where they may. As far as I understand, units will have their paths recomputed if they hit anything on their way - but I don't know how well it all would look. From what limited understanding I have, it should be fairly easy to simply cap the speed of the fastest unit to the speed of the slowest unit once the former has reached the latter. If for any reason the slow unit drifts out of range again, the fast unit would reset its speed to max until reaching the slow unit again.

I think this is the relevant piece of code: // Face towards the target if (!offset.IsZero()) { entity_angle_t angle = atan2_approx(offset.X, offset.Y); cmpPosition->TurnTo(angle); } In context, I interpret "target" as meaning "next waypoint on the path". If so, the call to cmpPosition->TurnTo(angle) would be what gives the abrupt, turn-on-a-dime behavior. Probably what happens is that the pathfinder traces a completely linear path for the unit and the UnitMotion component follows it slavishly, so jerkiness arises around the 'joints' of this path. If this is the case, I think fixing this would actually require changes to the pathfinder to have it create smoother (non-linear) paths. We can probably not just interpolate a curved path from the linear path, because an interpolated path might collide with obstructions that only the pathfinder can account for.

I guess it would be technically possible to do something like that without having to create new models, but it looks a bit too modern in this particular implementation (ancient folks wouldn't use 'building modules' like that). Perhaps someone skilled enough could find a way to derive a frame model from the basic building model, but it would be non-trivial (beyond non-trivial probably).

This is a trivilality, but is this intended: Seems like an OS X specific option?

I see. Yeah, I bet that looks fine.

That should be a really good quality assurance exercise, one that can be done on the forum: 1. what graphics card do you have? 2. does it work? 3. if no, let's try and fix it. 4. if it still doesn't work, we'll add it to the list of cards with known issues.

I saw this on Trac: If scaffolding is done (I realize it is hypothetical at the moment), presumably building frames would also need to be done? If frames are done, I wonder if we could also use them for rendering nice looking damaged buildings (like where part of the roof is gone, so you see part of the frame/interior instead)?

Looking forward to it!

I was actually wondering about that too. After watching a few youtube videos, my best judgement is that they do move/flicker noticeably. This is supposed to be recorded in real-time: (Edit: I guess the reason they are mostly seen in still-shots or time-lapse videos is that they need a long exposure time to look good on photos.)

It's fascinating how much is accomplished in just a week. I think up you can stir up quite a buzz with these reports if you don't burn yourself out from doing them so frequently. Consider doing them bi-weekly if you feel you are running out of material. Also, that pillar looks rad. Is it normal/parallax mapped? I hope it will be.

This is premature so I am posting it here Valve Software's Rich Geldreich will be giving a presentation entitled "Left 4 Dead 2 Linux: From 6 to 300 FPS in OpenGL" at this year's SIGGRAPH about custom profiling and optimization of OpenGL under Linux. They seem pretty serious, so it might be worth a look for anyone who has access...

I think it is less about being a boss and more about having an "agreement" or "mutual understanding". No one is going to want to work on a project where they can't count on the end result being something they deem worthwhile.

I suspect we are talking about different 'errors' The flickering is a known issue that can be attributed to existing code in the engine. The odd error messages you got ("preposterous datagram length" and all that), however, is not. So we are curious to know if these error messages happens with the vanilla GLSL renderer or only after you have applied myconid's changes.

Yes, it has been suggested repeatedly over the years. No one has bothered implementing it because it is not supported by the design document. I'll be counting the days until someone decides to do it just for fun

Do you get the errors on the vanilla build and preferglsl=true also, though?

Yeah, it seems kinds of elitarian, which doesn't speak to me either. Sadly, thinking that it is possible is unlikely to get us far.

Are there any historical examples of this or would it be purely fictional? If it would be purely fictional, how fictional would we allow it to be (if supernatural elements are out of the question)?

Not bad, but where does the player come into the picture? (Given that you'll have several players in a multiplayer game and they can't all be Alexandros IV ) In Diablo they used a "lone hero" plot device - you're a band of free-living adventurers / bounty hunters who become entangled in a war between Heaven and Hell. That way, the story would work no matter how many players you were (max 4) or which character classes you picked.

AFAIK, though wind adds the energy that indirectly creates waves, shore waves generally flows towards the shore, not in the wind direction. Non-shore waves are another matter, though.

It's not a bad idea, I admit, but there would still have to be some semblance of a story or we might as well strip the "RP" part from the ARPG moniker. What I liked in the Diablo series was the "chosen one" theme - this sense that the whole world was about to succumb to dark powers and you were the only one able to prevent that from happening. Not saying that we need to duplicate that plot, but some kind of plot would be good.

I certainly won't stand in the way of you or anyone else implementing these changes just for the heck of it. I just can't imagine myself doing it. Is there anything wrong in living documents? The main point would be to have an informal 'contract' between the various developers involved at a given time. If those projects didn't even have the cohension to produce a realistic design document, perhaps it's for the better that they didn't begin producing any code? I personally wouldn't mind dark fantasy staple 'supernatural' elements like spells and demons. It would, IMO, be invigorating to work on something completely orthogonal to the historistic constraints of 0 A.D. I agree that it should be easy for third-parties to add new 'quests' or 'missions', but I would definitely like a collected 'main' storyline too. There should be plenty of skilled writers on the web who would be capable of writing such a storyline - but someone would have engage those people, and that is not my area of expertise/motivation. Bottom-line: for a project like this succeed, it takes a certain amount of drive to get the right combination of talent together. If it was easy, video games wouldn't be a multibillion dollar industry.

I mean something like this. Something that shows the existence of a coherent vision that is worth pursuing. Specifically, I'd be interested to know about backstory, setting, storyline, gameplay/goals and needed art assets. Without at least a rough concept of these, we'll end up like Sumwars and the majority of other free software games:

I'll do it if I 'can see it happening' which means there has to be a design document of some kind.

Note to self (and anyone else who might be interested): source/simulation2/components/CCmpUnitMotion.cpp is probably the relevant source file.